DISCUSSION PAPER SERIES

IZA DP No. 11223

Alex BrysonRichard FreemanRafael GomezPaul Willman

The Twin Track Model of Employee Voice: An Anglo-American Perspective on Union Decline and the Rise of Alternative Forms of Voice

DECEMBER 2017

Any opinions expressed in this paper are those of the author(s) and not those of IZA. Research published in this series may include views on policy, but IZA takes no institutional policy positions. The IZA research network is committed to the IZA Guiding Principles of Research Integrity.The IZA Institute of Labor Economics is an independent economic research institute that conducts research in labor economics and offers evidence-based policy advice on labor market issues. Supported by the Deutsche Post Foundation, IZA runs the world’s largest network of economists, whose research aims to provide answers to the global labor market challenges of our time. Our key objective is to build bridges between academic research, policymakers and society.IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character. A revised version may be available directly from the author.

Schaumburg-Lippe-Straße 5–953113 Bonn, Germany

Phone: +49-228-3894-0Email: publications@iza.org www.iza.org

IZA – Institute of Labor Economics

DISCUSSION PAPER SERIES

IZA DP No. 11223

The Twin Track Model of Employee Voice: An Anglo-American Perspective on Union Decline and the Rise of Alternative Forms of Voice

DECEMBER 2017

Alex BrysonUniversity College London and IZA

Richard FreemanHarvard University and IZA

Rafael GomezUniversity of Toronto

Paul WillmanLondon School of Economics

ABSTRACT

IZA DP No. 11223 DECEMBER 2017

The Twin Track Model of Employee Voice: An Anglo-American Perspective on Union Decline and the Rise of Alternative Forms of Voice

We present a simple framework for analyzing decline in union voice in the Anglo-American

world and its replacement by non-union, often direct, forms of worker voice. We argue that

it is a decline in the in-flow to unionisation among employers and workers, rather than an

increase in the outflow rate, that accounts for this decline. We show how union decline

is predicted by experience good and cost-disease models of trade unionism and is linked

to specific institutional and policy constraints on union organizing in the Anglo-American

world. We show how the coexistence of union and non-union forms of worker voice is

predicted by transaction cost economics, while the growth in non-union forms of worker

voice is aided by declining costs of employers “making” voice mechanisms. We draw on

“spurt” theories of unionisation to help understand factors underpinning union decline,

including falling costs of employer opposition to unionisation as density falls, as discuss

possibilities for “bottom-up” growth in union-like forms of worker voice implied by “spurt”

theories.

JEL Classification: J51, J52, J53, M54

Keywords: worker voice, unionisation, collective bargaining, experience good, cost-disease, transaction costs

Corresponding author:Alex BrysonDepartment of Social ScienceUniversity College LondonGower StreetLondon WC1E 6BTUnited Kingdom

E-mail: a.bryson@ucl.ac.uk

2

1. INTRODUCTION The effect of trade unions on the outcomes of firms and workers has been a core concern of

labour and industrial relations (IR) scholarship for as long as IR has been considered a distinct

field of study (Webbs, 1897; Dunlop, 1944). Indeed, some labour historians have gone so far as

to argue that the debate over the “labour problem” that emerged in the wake of the Industrial

Revolution was instrumental in the birth of the new science of IR during the late19th and early

20th century (Kaufman, 2010).1 Essentially, industrial relations experts were “in demand” to

solve real tractable problems like absenteeism, sit-down strikes and violent clashes between

workers and management (Derber, 1967). One of IR scholarship’s earliest recommendations was

for workers to have some mechanism for collective representation at the workplace (Feldman,

1928) and independent trade unions were seen as the closest embodiment of this goal

(Commons, 1935).

Although labour scholarship no longer uses terms like “industrial man” or the “labour

problem” (Dunlop, 1958) to frame the debate over unions and the labor market, current concerns

about the causes and effects of trade union decline are remarkably similar to those of nearly a

century ago. In particular, the problem of fostering industrial democracy, as it [i.e., employee

voice and representation] was originally termed by scholars like the Webbs (1897) and

Commons (1935) was still at the forefront of research and policy making well into the 1970s

1 A reasonable argument could also be made that “scientific management” and industrial engineering (IE) was an equally compelling creation of the industrial revolution in that industrial engineers provided a ready-made alternative solution to IR experts and unions in resolving the difficulty of managing large numbers of laborers (Jacoby, 1985, 2004).

3

(Milton, 1970). Now after a long period of dormancy, the issues of worker representation and

employee voice are making a noticeable return in public policy debates. 2

As seen in Figure 1, the main problem facing scholarship in the area of employee voice,

especially in the Anglo-American world,3 is that the main vehicle for providing voice and

representation to workers (i.e., the trade union) has witnessed a near universal decline over the

past 40 years. This raises the inevitable question of just how important voice and representation

is in the context of post-modern/post-Fordist economies (Thompson and Newsome, 2004).

[Figure 1]

We begin by presenting a simple framework for analyzing union decline that follows

some of our own work in this area (Freeman, 1998; Freeman and Rogers, 1999; Bryson and

Gomez, 2005; Willman, Bryson and Gomez, 2006, 2014; Bryson, Willman, and Forth, 2017).

We model the decline in union representation as a shift in uptake by workers and firms between

(at least) two types of employee voice –from representative (collective) to direct (individual)

voice. We discuss the potential channels that link a rise in “direct (individual) voice” and a

commensurate fall in “representative (collective) voice” to specific subgroups of firms and

workers, focusing on two alternative hypotheses: (1) the loss of unionised jobs and a decline in

2 The 2016 British government green paper on Corporate Governance Reform and Theresa May’s recent calls for worker representation on public company directorships suggests that this new-found interest in industrial democracy will remain on the political agenda as an element in wider reforms to boost social cohesion, both in Britain and internationally. Many countries currently face similar social strains and are looking to a revitalised workforce as an answer. Some, like the Trump administration and Brexit advocates, are trying to improve domestic wages through a combination of immigration restrictions and managed (as opposed to ‘free’) trade. A longer-lasting and empirically successful alternative would be a Nordic-model of economic openness coupled with a high degree of worker protection and collective bargaining coverage. 3 We could of course add countries like Australia, New Zealand and Ireland to the mix but then the term ‘Anglo-Saxon’ would be ill-fitting for a country like Ireland often billed as the ‘Celtic tiger’. Perhaps a more neutral phrasing like English speaking countries works though even here we would need to qualify that we are focusing on advanced/industrialised economies as well. This is why we settle on the Anglo-American world as our country-focus descriptor.

4

union density in established firms and amongst older workers; versus (2) new workers and new

firms adopting alternative forms of non-union voice or no voice at all.

We show that it is more the latter channel, the rise of so-called ‘never-membership’ and

rise of alternative voice systems in newly set-up firms, which has contributed to most of the

decline in union density (especially in Britain where we have the best and longest time series

data on workplace voice).4

We then turn to a discussion of how the decentralised form of labour relations that

prevails in the Anglo-American world – specifically the Wagner-Act model with its emphasis on

organizing individual workplaces –not only accounts for the never-membership phenomenon but

has also fostered a “twin track” mode of employee voice whereby established sectors and

occupations with high barriers to entry -- e.g., where there are stable product markets or service

delivery monopolies (i.e., public services)-- enjoy higher rates of union voice, whilst a growing

new segment of workers and private sector firms exist as either “no voice” enclaves or have

filled the union-only voice void with employer provided systems and increasingly, it seems, with

self-organized on-line forums and social-justice campaigns.

Speculation of what lies down the road comprises the final section. We think some

traditional notions of voice as proposed by authors like Barbash (1989), Meltz (1989) and Budd

(2004) have salience in this new era since their basic thrust was that a “law of equilibrium” or

“balance” between equity, voice and efficiency exists in the labour market. When the provision

of voice and equity is missing, different actors step in to fill the void. If all three traditional

4 A study of never unionism by Booth, Budd and Munday (2010) finds never-unionization rates in the United States are also increasing as new cohorts enter the labour market, even after adjusting for demographic and structural changes. In the mid-1980s, 58.1 per cent of 23-year olds had never been unionized, but roughly 20 years later, 70.6 per cent of 23-year-olds had never been unionized.

5

actors are absent (i.e., management, government, unions), this is when we would expect new

forms of voice propelled by new actors to emerge.

2. DECENTRALISED BARGAINING AND THE RISE OF TWIN TRACK UNIONISM The early 1980s marked the high-water point for union representation across the Anglo-

American world (see Figure 1). In 1980, for example, five out of ten workers (51 percent) were

represented by trade unions in Britain. By 2014 that number had fallen to just one in four

workers (25 percent). In the United States representation was always on the low end of the cross-

national scale (see bottom line in Figure 1), but more disquietingly perhaps, is that this overall

decline masked an even sharper fall in private sector union representation, something that has

afflicted all Anglo-American countries. Whereas union density rose in the public sector during

the late post war period in the United States, in the private sector it has been falling continuously

since as far back as the 1960s (Figure 2). The same pattern of high public sector density and

declining private sector membership since the 1970s prevails in Britain and Canada.

[Figure 2]

The three panels for the US, Canada and Britain look very much like unionism has been

on a “twin track” for some time, with a sectoral bifurcation that is only one of many that exist in

the provision of union voice across all Anglo-American countries. Given such a marked decline

in such a vital labour market institution in the private sector, several questions arise.

• First and foremost is the question of why? Why such a decline in collective representation

almost entirely concentrated amongst private sector workers?

• Second, does this decline call into question the relevance of trade unionism or the need

for any other form of employee voice?

6

• Lastly, in light of past practice, current industrial and workplace make up, and more

pointedly, given the shifting nature of work and employment under the auspices of what

has come to be known as the ‘4th industrial revolution’, ‘gig economy’ or ‘new world of

work’, what forms (if any) of collective representation and voice might be possible in the

Anglo-American world?

To answer these questions, we need to first examine the Anglo-American model by which

employees secure union representation and voice. Though each country differs, the Anglo-

American model nevertheless contains a set of features that are common to all systems where

certified bargaining agents acquire exclusive representation rights and where the employer

has no statutory obligation to deal with sectoral bargaining agents or minority unions.5 In

this system -- sometimes termed “exclusive representation with majority support” or simply

the Wagner Act Model (WAM) -- workers who have certified bargaining agents are

commonly referred to as “unionized” or “organized” while those with no certified agent are

referred to as “non-unionized” or “unorganized”. Though different mechanisms exist across

Anglo-American economies for securing trade union representation (e.g., some jurisdictions

invoke easier certification processes while others prevent agency-provisions)6, the Anglo-

5 In Canada and Britain, unlike the US, there is a tradition of “voluntary recognition” in which an uncertified employee association can bargain on behalf of its members with management, so long as management willingly accepts. Such employee organizations sign memorandum of agreements (MOUs) with the employer and confer a considerable range of protection for workers. They can also in Canada finance their operations through automatic dues payments (agency shop provisions) much in the same way as certified unions do (Taras, 2006). For the most part, however, Canada unlike Britain, operates like our southern neighbour and representation is secured through formal certification procedures. 6 Two systems exist for obtaining union certification in North America: The Card Majority Certification Regime (card system) and the Mandatory Secret Ballot Vote Certification Regime (vote system). Under both systems, the first stage in applying for union certification is proof of support for the trade union in a bargaining unit. Under the Card Majority Certification Regime, adopted by four of the ten provinces in Canada (Quebec, New Brunswick, Manitoba and P.E.I) union certification will succeed if a majority of employees represented by the bargaining unit are signed up. This majority varies among the different jurisdictions: New Brunswick and Manitoba require a super

7

American model in its essential form displays the following seven characteristics that are

present in all workplace-based majority representation systems:7

a. “Seven” Key Features of Decentralised Representation Systems in the Anglo-American World

i) A non-union / no-voice default setting for employees – i.e., whether working in a bank

branch or retail outlet there is no system of worker representation in nascent organizations. That

is, private sector workplaces are set up and largely exist in an entirely unorganized,

unrepresented, non-union condition. As a result, private sector employees, for the most part,

work in non-union workplaces and have no access to formal representation systems.

ii) If an employee has access to some modicum of voice and representation within the

workplace, it is because the majority of workers were first organized to undertake the process of

joining an independent union, securing certification from the labour board and then signing a

first collective agreement with the employer. For the typical worker, possessing no past

experience with unions or certification procedures, this process can appear lengthy and

complicated. In cases where employers are opposed to unionisation, the process is tilted against

workers gaining certification (Doorey, 2013). In particular, the use of mandatory vote

certification procedures (as opposed to card-based certification which exists across some

Canadian jurisdictions) has had a negative effect on certification success rates and on union

density (Johnson 2002, 2004). Slinn and Hurd (2011) argue that in Canada certification delays

often arise from unfair labour practices (ULP) applications and hearings related to employer

majority of signed cards (60%+1 and 65% respectively), while a simple majority (50%+1) is sufficient for automatic certification in Quebec and Prince Edward Island. Even under this system, should a group not obtain automatic certification; labour law allows a vote when another threshold of signed cards (obviously lower than for automatic certification) is met. In the US it is exclusively a vote system that prevails nationally and in Britain unions can choose long-standing voluntary recognition or a certification vote similar to the North American model. 7 This is a list adapted from an original list of five features of the US system found in Weiler (1989).

8

conduct during organizing while in the US a similar pattern prevails (Ferguson, 2008).

Moreover, there is evidence that employer ULPs are substantially more effective at defeating

unionization under mandatory elections than under card check certification (Riddell, 2001).

Though British employees face similar hurdles in organizing their employer as do their

North American counterparts, getting to first contract is not an issue in the same way because

although in law a collective agreement can be directly binding, it is usually (nearly always) the

case that collectively agreed upon terms are implied terms in individuals’ contracts of

employment. Thus, a collective agreement is almost never directly legally enforced.

iii) Non-union employers can be quite averse to unionisation drives and as such deploy a

variety of tactics in order to head off incipient unionism (Riddell, 2001; Slinn, 2008; Doorey

2012). Though these tactics have been associated more closely with US style labour relations

(Logan, 2006), comparative research has actually shown that managers in other countries report

equal, if not greater, negative attitudes towards unions than their American counterparts

(Beaumont, 1986, 1994; Campolieti et al., 2007, 2013; Lipset and Meltz, 1997; Saporta and

Lincoln, 1995; Taras, 1997). Many non-union employers also engage in sustained efforts to

persuade and supplant independent trade unions while others will coerce and supress incipient

unionism (Thomason and Pozzebon, 1998). Furthermore, although there is evidence that major

companies headquartered outside of the US have a lower likelihood of opposing unions with the

same kind of zeal as US counterparts (Thompson, 2001), union acceptance amongst employers

in Canada and elsewhere, as Rose and Chaison (1996: 92) presciently observe, ‘is most likely the

result of a low probability of escaping unions rather than a [less inherent antagonism]’. There is

also a particular UK employer response to unions, found in recent analyses of WERS data in

Britain, which can perhaps be best couched as “apathy”; a recognition perhaps that organizing a

9

workplace is difficult and likely to fail on its own terms such that expending valuable resources

in opposing a union is not only unnecessary but possibly counterproductive (Amosse, Bryson,

Forth, and Petit, 2016).

iv) As a result of the difficulties in establishing union representation – as documented in

i) to iii) -- unionisation can be conceived of as an ‘experience-good’ which is defined by Nelson

(1970) as any good or service whose quality cannot be truly discerned before purchase. For

workers (and employers alike) unionisation is an experience good in the sense that most union

benefits (procedural justice, job security, the provision of family-friendly policies) are hard to

observe before joining a union or working in an organized workplace. Experience goods in turn

have certain properties that make them hard to ‘market’ to potential adopters who have never

sampled membership. In our case this means that groups like young workers or immigrants are

particularly likely to by-pass union organizing unless otherwise compelled to do so (Gomez and

Gunderson, 2004). Moreover, even if attributes are made visible (through information

campaigns) union provided benefits are still of indeterminate quality to a non-union worker (i.e.,

“you don’t know how good a union is until after you join”) and hence still subject to the same

experiential marketing challenges.8 The indeterminate nature of benefits associated with

unionisation can generate hesitation and skepticism on the part of non-union workers, which

means that organizing to join a union is difficult since it is a decision whose payoff is only fully

revealed after joining.

v) Most workers are members of large (more than 100,000 members) unions. This is

natural given that in order to establish a critical degree of employee interest and to overcome the

sustained resistance of some employers, the attempt to provide voice and representation rarely

8 This is why some unions in the UK and Australia (e.g., the Independent Education Union (IEU)) have experimented with “free” membership for students and newly hired teachers (Costa et al., 1997).

10

succeeds on its own. It is instead necessarily taken on by large national (or international) unions

that have ample resources to overcome the fixed costs associated with new certification drives

and experience organizing under the Wagner model in order to lower the marginal costs

associated with representing new workers. Large national/international unions are less

successful, however, at unionising new firms and small-to-medium sized enterprises in the

private sector; instead they deploy scarce organizing funds to certify large well-established firms

with large bargaining units (Willman, 2001). Rightly or wrongly, larger unions are also

perceived by many non-union employees as having a distinct impersonal and bureaucratic

flavour (Weiler, 1989).

vi) Partly as a result of i) through v) there has been a steady decline in unionisation in the

private sector and a growth in “frustrated demand” for employee representation that is not being

met (Freeman and Rogers, 1999). Frustrated demand for unionisation is generally greater for

younger workers. For instance, Bryson et al., (2005: 165) found youth (<25 years) representation

gaps of 36 percent (50–16 percent) in Britain, 44 percent (57–13 percent) in Canada, and 42

percent (47–5 percent) in the U.S. The corresponding adult (25-64 years) representation gaps

were 11 percent (46–35 percent), 12 percent (48–36 percent), and 37 percent (53–16 percent),

respectively.

vii) This large representation “gap” has been filled, in part, by a growing array of

statutory provisions (e.g., minimum notice periods) and regulatory mechanisms (e.g., minimum

wage laws). Unfortunately, as pointed out by legal and economic scholarship, these universal

regulatory standards are subject to the problem of “unfulfilled legal promise” (Weil, 2007).

Unless there is an indigenous base of union representation or workplace oversight, the internal

11

vigilance required to make the universal employment standards programme a reality is often

lacking in non-union workplaces and compliance is attenuated (Barbash, 1987; Meltz, 1989).

Given the problems associated with decentralised majority-based collective bargaining

outlined above, the rest of the paper identifies an analytical framework to understand union

decline in Anglo-American economies and ends with a discussion of the future of voice. The

series of employee/workplace voice options for the unrepresented worker and unorganized

workplace cover the gamut of employment relationships: from employees working in traditional

workplaces under standard employment relationships (e.g., full-time work with permanent as

opposed to temporary contractual arrangements) to employees lacking a clearly defined

employer (e.g., temp- agency workers) or workplace (e.g., freelance workers).

A common thread linking all possible voice ‘futures’ is a resetting of the current default,

whereby employers (and their associated workplaces) are “voice free” until such time as an

indigenous non-union scheme is established or an independent trade union emerges through a

formal certification process (Doorey, 2013; Braley-Rattai, 2013). An alternative is to fashion a

system akin to health and safety whereby all employers are obliged to conform to some

minimum requirement for employee consultation and participation. The particulars of which,

depending on their size and nature, would be enshrined in standards that apply to all workplaces.

Or perhaps consideration should be given to a “minority union model” whereby in an appropriate

bargaining unit, the most representative union (or coalition of unions) with a minimum

membership to make it credible, could be certified by a labour relations board as a primary

bargaining agent with all of the rights and duties of exclusive agents (Adams, 2015).

Underpinning these and other options is a broader view, held by many in the employment

relations field, that every worker should as a matter of principle be afforded some system of

12

employee voice (Adams, 2008; Budd, 2004) irrespective of whether this form of voice has been

adopted by the firm or whether a union representation system has been secured via the

certification process common across most of the Anglo-American world (Slinn, 2008; Doorey,

2013). And drawing from the earliest scholarship in our field, it should be noted that while the

‘under-provision’ of collective voice may be optimal for an individual employer wishing to

secure employee compliance without investing in commitment mechanisms, these private

choices may be suboptimal from a societal perspective, either in terms of low economy-wide

productivity emanating from lost shop-floor knowledge or from a lack of meaningful

participation in an industrial democratic sense.

3. A FRAMEWORK FOR UNDERSTANDING UNION DECLINE AND THE RISE OF NON-UNION VOICE

We argue that the fall in private sector unionism and the high and relatively stable rates in the

public sector can be accounted for by a model of unionism where union growth (or decline) is

part of a “bottom-up” process (Freeman, 1998) embedded in a wider set of competing and

complementary voice channels – all of which exist in a solution market for employee voice.9

Each voice channel – union, non-union or some dual form hybrid – requires a set-up ‘cost’ and

has to be experienced first in order to discern true quality and the full scope of benefits. This idea

is captured by the “experience-good model of employee voice” (Bryson and Gomez, 2003, 2005;

Gomez and Gunderson, 2004; Budd, 2010) which borrows some basic elements of consumer

9 In the union case it is arguable that the ‘experience good’ in question has three components: 1). A right to have a say over the conditions of work and pay (i.e., voice); 2). Some measure of bargaining power such that a greater share of profits flowing to workers than would otherwise be the case (i.e., wage premium); 3). Some modicum of job security provided via internal grievance procedures and just cause termination provisions (i.e., insurance). All three elements are only revealed to workers after they have secured representation, or in the parlance of the model, have “purchased” membership

13

theory to predict the way in which voice provision differs amongst groups and how voice

provision has shifted over time.

We highlight evidence that is potentially consistent with the experience-good hypothesis

and those that pose either a problem or a puzzle for the model. By problems, we mean the

evidence suggesting that workers desire union representation but are not being organized by

labour unions in sufficient numbers to increase density figures, which at least superficially

appears to be inconsistent with the view that workers ‘truly’ desire union alternatives over

employer voice.10 Troy (1999) and other authors have used this fact to advance the claim that

collective representation has simply been overtaken by self-representation and individualised

contracting as the preferred employee mode of employment.

We disagree with this reading and instead rely on the ‘incumbency effect’ (Freeman and

Diamond, 2001), which states that any established employee voice system (provided it meets a

minimum standard of acceptability) will be favoured over a proposed alternative, even if that

alternative is better in some objective way. The acquisition of switching-costs for workers (and

firms) is one of the key elements in experience good-theory that ties the incumbency and

experience-good models together such that an incumbency effect is captured by the added “cost”

and/or “risk” of switching out of one employee voice system to another of lesser-known quality

(Willman et al, 2014). Switching costs can keep employees locked-into sub-optimal employment

10 Bryson and Freeman (2013) find a correlation between increasing problems at work and the desire for ‘solutions’, and when the desired solution was more union-like, workers seemed to gravitate to that suggesting that there is clear evidence of a representation gap. But as with any desire (I may want to drive a lambourghini) there are costs that workers have to ‘cross’ in order to be unionised and relative to benefits those costs may be too high. And second, perhaps the defining factor is the reduced desire for union voice on the part of employers. Indeed it is, arguably, this rather than changes to employee desire that is deterministic since in the Anglo-Saxon world unions require an employer 'host'. The decline in unionisation amongst new workers is really just a reflection of their concentration in new firms which have chosen not to unionise for the reasons explained in our model.

14

relationships just as they can keep consumers locked-into sub-optimal mobile service contracts

or bad banking relationships.

By puzzles, we mean evidence that is potentially consistent with our version of

experience good theory but appears to be driven by other causes. For example, there is increasing

evidence that the standard labour organization itself– i.e., the modern trade union – has

significant “cost-disease” attributes (Willman, Bryson and Forth, 2017) that in the current

legislative and economic climate render their operation highly tenuous and which might explain

their current lack of large-scale organizing success despite an apparent appetite amongst many

workers for representation.

Our conclusion is that the experience-good hypothesis is an accurate depiction of why

unionism has slowly ebbed as the default form of voice in Anglo-American economies but falls

short as a mono-causal explanation for the decline of unions and evolution of alternate forms of

voice in the most recent period of the early 2000s to the present. A theory compatible with

experience-good unionism is one which sees the emergence (or decline) of unions and worker

voice systems as part of a “bottom-up process” in which individual actors make decisions, adopt

new strategies and/or engage in behaviours that can (cumulatively) generate sharp changes in

trends or likewise contribute to long periods of union stagnation or decline (Freeman, 1998).

Though it seems like union decline has been the dominant theme since the 1980s, this is

not to say that we believe traditional unions have become irrelevant, or that recent changes to

their organizing strategies have had no effect on certain groups of previously unorganized

workers (i.e., organizing public sector workers in the 1960s and 1970s or extending their reach

through Justice for Janitors campaigns in the 1990s etc.). There were clearly innovations in union

organizing over the past 30 to 40 years and it seems likely that these changes had some effect on

15

new groups of workers (Erickson et al., 2002). Rather, we argue that while existing models can

account for the presence of a pretty stable (albeit declining in relative terms) group of workers

and firms that have union representation and therefore have access to collective voice; industrial

relations theory, up to now, has not been particularly helpful in understanding the shifts in the

structure of new forms of voice and the birth of “do-it-yourself” organizing that has occurred in

Anglo-American labor markets over the last decade.

Based on our reading of the evidence, we believe it is time to re-evaluate the rise of

individualised and increasingly digitally infused forms of worker voice that have emerged in the

first quarter of the 21st century.11

We begin, however, with a look as to why union voice has declined, focusing on the

proximate cause first, namely the rise in the share of workers that have never sampled union

members or have never been covered by a collective agreement (i.e., ‘never-members’).

a. The Rise of the “Never- Member”

Though many authors have contributed to the scholarship on union decline and noting that the

reasons for union decline in the Anglo-American world are multi-causal (reflecting changes in

industrial composition, free-trade and the mobility of capital, low inflation and high

unemployment, conservative governments and associated legislative swings etc.,), we wish to

11 The difficulty in this area is that it is very hard to move beyond the few anecdotal case studies that exist. These new employee advocacy groups are so new that they still have not appeared in a systematic way on many nationally representative survey designs. Specifically, the question becomes how does one locate these new forms in the context of the theory and models discussed above? We know in theory that they can reduce the union cost disease problem by 'farming out' union organising and services to committed volunteers – the so-called 'off balance sheet' work of unions. But perhaps more importantly there appears to be a fluidity to the organisational form that these on-line union tools take which means they do not conform to standard models of trade union representation which pivot around collective bargaining rights at a specific workplace. In militarsitic terms the analogy we have in mind is conventional ground forces (old unions) versus guerilla forces (new unions).

16

focus on a finding first observed in Britain. We do this in order to point out that irrespective of

underlying causes, the proximate cause of union decline has more to do with several features of

Wagner-style representation systems highlighted above than anything else.

Bryson and Gomez (2005) were the first to track the rise in the percentage of employees

in Britain who had never become union members (‘never-members’) since the mid-1980s (see

Figure 3). The data for Britain showed that never-membership had increased markedly in the 20

year period between 1986 and 2006 whilst ex-membership had remained roughly constant. The

work also showed that by the early 2000s, never joining a union had grown steadily for the

average 30-year old worker outstripping changes in other major life-course events such as

marriage and university attendance. Bryson and Gomez (2005) went on to show that it was this

reduced likelihood of ever becoming a member and not the loss of existing members which

accounted for the decline in overall union membership in Britain since the 1980s.

[Figure 3]

In the United States, though we do not have a comparable never membership figure, we

know from Budd and Booth’s (2010:42) analysis of successive waves of NLSY data that it “does

indeed appear that never-unionization rates in the United States are increasing as new cohorts

enter the labour market, even after adjusting for demographic and structural changes.” Such a

view is also consistent with what has occurred in a Canadian context. Since 1997 Canada's

unionized workforce has actually grown in absolute terms. Presently, roughly 4.56 million

workers are members of unions — or are covered by collective agreements or union contracts —

up more than 800,000 since 1997. We also know that de-certification in Wagner-style systems

17

offering first-contract arbitration is rather rare (Baker, 2012).12 So why then has union density

declined across Canada?

The reason is that union membership expansion has been far outstripped by the growth of

the non-unionized workforce, which grew by about 2.5 million over the same period. In other

words, new workers are not being organized at a sufficient rate to increase union representation

as a share of the overall workforce. Put simply, never-membership is rising in Canada as it has in

Britain and in the United States and this has negative implications for the future of the Wagner

Act Model (WAM) of representation (Booth, Budd, and Munday, 2010).

The rise in never-membership, however, is still a by-product of some set of larger

underlying processes and as such is still only the “proximate” cause of union decline in Anglo-

American economies. To understand the deeper causal processes at work we need a more

‘general theory’ of union growth (and decline).

b. The “Spurt Theory” of Union Growth and the Rise of Right-to-Work Laws

As noted by Freeman (1998) two types of models can generate changes in union growth. The

first, and probably most often used in IR scholarship, are models in which massive shocks or

environmental changes generate commensurately large responses in otherwise stable union

membership. The second are models in which the process of growth creates nonlinearities that

produce “phase transitions” when certain conditions are met, generating models of tipping-point,

contagion, self-organized complexity and so on.

12 Unless a workplace closes or moves, a union voted in by a majority of the workers in Canadian jurisdictions tends to stay in place. As noted by Baker (2012) “Unlike the US, most provinces [such as Ontario] have laws that provide for first contact arbitration in the event of a deadlock in negotiations. Almost half of all successful unionisation drives in the US do not lead to a contract. While the company is legally obligated to negotiate in good faith, this is generally not much of a requirement. Delaying a first contract is an effective way to undermine support for a union and often leads to a union being decertified. Provinces with first contract arbitration have higher rates of unionisation [and lower de-certifications].”

18

The latter models stress the underlying process by which organization occurs and the

cumulative effect of individual actions taken by workers, unions and firms. The focus is on the

behavior of many individual actors acting in response to one another. The former models stress

the exogenous shock, usually generated by political forces which is perhaps why legal scholars

and labour historians gravitate to these explanations and generally interpret the growth or decline

of unionism in their respective country as resulting from one (or more) of these “one-off”

legislative changes: i.e., unions declined in the United States because of Reagan’s firing of

PATCO strikers, in Britain they fell because of Thatcher’s crushing of the coal-miners, in

Canada they grew because of the enactment of PC 1003, and so on.

Without denying the importance of particular laws or external events as catalysts for the

growth process, Freeman (1998) lays out a model in which sudden sharp changes in union

density (what he calls ‘spurts’) do not depend upon “external environmental changes” but rather

arise endogenously from the process of union organization.

There is an affinity between Freeman’s spurt model and a related theory of industrial

mobilisation proposed by Kelly (1998), which generates a similar tendency for explosions in

industrial militancy and union growth. These shift-changes coincide with critical turning points

in “long wave” political-economic developments. Kelly’s use of mobilisation theory reveals the

problems faced by workers in pursuing a course of collective action to settle grievances and

promote their own goals. In particular, he emphasises the importance of workers acquiring a

collective awareness of injustice emanating from the social relationship between capital and

labour, where employers and employees pursue different and/or opposing sets of interests. The

question is how can this collective awareness emerge or be ‘mobilised’? Mobilisation, according

to Kelly, depends on the accessibility of effective worker organisations and the availability of

19

labour leaders willing and able to head workers' struggles. The opportunity to take collective

action, however, can be curtailed by employers via sticks (threats such as firings and plant

closures) or carrots (inducements like profit sharing or internal grievance procedures). The state

can also enter the labour-capital relationship typically (though not always) on the side of

employers with legislation and actions that thwart worker organizing efforts.

What’s interesting is that Freeman’s (1998) model identifies the same forces of union

‘mobilisation’ (organizing) and ‘employer threats’ (opposition) identified in Kelly’s theory of

discontinuous union growth and decline, but formalises them into union “organizing” and

employer “opposition” behavioural functions. Freeman’s model can explain and anticipate the

nonlinearities present in union density over time and across space as functions of divergent union

and employer behaviours. As such, we see the two theories (‘mobilisation’ and ‘spurt’ theory) as

complimentary to each other. The key to Freeman’s model, however, is that employer opposition

is linear in that anti-union efforts decline as union density rises while union organizing is non-

linear (i.e., weakest at the lowest and highest union density levels). The reason is that at very low

density levels unions lack resources to mobilise, whereas at high levels, there is little incentive

for unions to organize given that the small number of non-union workers have little incentive to

join given positive union-spillover effects. The union organizing function therefore rises and

falls in an inverted-U fashion with increasing union density.

There are two stable equilibria in the Freeman model, a zero union density equilibrium

(though Freeman acknowledges that in practice there would naturally be some unionism at the

theoretical ‘zero’ mark due to the difficulty in removing unions where they are present) and a

high (though not 100 percent) union density equilibrium. A low (non-zero) union density level

generates instability in a positive or negative direction for unionisation. Moves in a negative

20

direction are initiated by employers given their opposition to unions increases as density declines

(i.e., no employer wants to be the single employer encumbered by a collective agreement or

having to pay a union premium). Given the difficulties in organizing at the best of times, moves

in a positive direction from a low union starting point generally coincide with a confluence of

external factors but more importantly with union innovations in organizing that appear to shift

the organizing function upward. The well-noted “strategic-choice” approach advocated by

Kochan, Katz and McKersie (1983) to explain how differences in labour-management outcomes

can emerge in firms facing observably similar external environments is therefore compatible

with Freeman’s “spurt theory” of union growth. Innovations can emerge locally from actors

facing otherwise similar external constraints but it is only when they spread in quick succession

to others that innovations in worker organizing become transformative.

The empirical evidence highlighted by Freeman (1998) shows that all spurts in union

density were generated by every union adding members. However, a much larger proportion of

the growth in the 1930s and 1940s (the largest era of union growth in the US) occurred through

“new” CIO industrial unions emerging and organizing new workers in new ways than in previous

spurts, which explains the uniquely large growth experienced in that period.

The spurt in public sector unionism that occurred in North America in the 1960s and

1970s had properties similar to the private sector growth noted above (i.e., rapid expansion of

unionism into new areas with new or changed organization forms such as the switch of employee

associations like the National Education Association [NEA] into certified unions) and growth of

existing unions under the help of legal changes (Freeman and Ichniowski, 1988). It is interesting

to note that the levels of unionisation achieved in the public sector across most of the Anglo-

American world (and as noted in Figure 2) have been high and fairly stable. It might be that they

21

have reached (for a given level of employer opposition in each national context) the high-level

equilibrium anticipated in Freeman’s model. In Britain, US and Canada the public sector

unionisation rate is roughly five-times the rate present in the private sector and the gap has

persisted for over 40 years in the case of the US and Canada and even longer in the case of

Britain.

One final striking fact anticipated by Freeman’s model requires mentioning: given that

bottom-up processes rather than external shocks are implicated in long-run changes in union

density one would expect that employer (and or State) opposition to unions is triggered when

organized labour is already weak and in decline rather than when unions are strong and on the

ascent. Such a tendency is revealed when we look at the US in more detail especially since the

1960s which coincided with the peak of the US union movement in terms of membership

growth. It was a decade or more after this peak that decline in density had begun in earnest and

that individual states began enacting right-to-work (RTW) laws, which limited trade unions

ability to collect dues from workers and, it is alleged, hampered unions’ ability to organize. The

passing of the Taft-Hartley act in 1947 led to an immediate adoption of RTW laws by

historically low-density Southern states but then after a thirty year period of stasis not a single

State enacted a RTW law. That is until 1976 whereupon states began enacting RTW laws such

that by 2017 more states had RTW laws than did not.

We compiled data (shown in Table 1) on when states (in the post 1960s time period)

enacted RTW laws and compared union density at the time of the law passing. In each and every

case, a RTW law was passed when union density was below (and well below in most cases)

historical highs. Never was it the case that a RTW law was enacted at a time when union density

22

was higher than its historical average, rather laws “restricting” union organizing were passed

when density was at or near the state’s historical low.

[Table 1]

What can we infer from these data? As it relates to the “spurt” theory of union growth the

answer becomes apparent once we visualise the data contained in Table 1. Graphically depicting

the emergence of RTW from 1964 to present, we get a feel for what happened from the peak of

the labour movement in the US to its nadir today. As shown in Figure 4 the first four states that

enacted RTW laws were Louisiana (1976), Idaho (1985), Texas (1993), and Oklahoma (2001).

All of these four states share the distinction of occupying the bottom four lines in the graph; that

is, they were the “lowest density states” and were below the US average at the time they enacted

RTW laws. In fact in all cases but one – Michigan – US states enact a RTW law only after state

union density falls below the US average (dotted line). This is a remarkable finding because it

supports Freeman’s model of bottom-up processes rather than external “shocks” through “legal

changes” as leading the decline of unions in the American context. Recall as well that the union

“opposition” function in Freeman’s model predicts just such behaviour in that union opposition

is inversely related to union density e.g., (low density engenders higher opposition). Laws and

legal changes are more like “trailing indicators” of organizational or movement

strength/weakness rather than generators of them.

[Figure 4]

c. The Experience Good Model of Union Membership and Implications for Employee Voice

The rise in never-membership, decline in overall density and the spurt model of union growth

documented above have several important implications for the future of traditional collective

representation.

23

First, when unions are present in an establishment they extend the benefits of union

representation to a wider set of workers than just the ones they represent. Whether it is through

the “threat effect” (i.e., the inclination of non-union employers to match or even supersede

certain contractually bargained outcomes in an effort to forestall union interest) or by lobbying

government for improved legislated outcomes, a strong union movement also affects labour

conditions outside of the organized sector.

Second, unions are successful in organizing workplaces when workers possess

considerable knowledge of what unions can do through mobilisation (Kelly, 1998) or when there

is strong demand for collective representation amongst workers (enough to overcome the

significant hurdles associated with majoritarian representation). As a result, any decline in the

number of workers who have ever sampled unionism makes the task of organizing individual

workplaces especially hard.

Third, to the extent that unionism is an “experience good” in the sense that workers can

only fully appreciate the benefits of unionization by experiencing it firsthand, or indirectly

through familial and collegial relations with other union members, then worker and sector

composition matters. If the bulk of workers who have never-sampled union membership are

concentrated in new sectors of the economy and amongst new workers, it is hard to see how the

majority-based workplace model-- designed as it was to organize workers with some knowledge

of how unionism functions and how certification is achieved -- can ever recover in a private

sector dominated increasingly by new (often hostile or apathetic) employers and new (often

uninformed) workers.

In short, gradual union decline is more likely when the default representation system in

place is one in which workers have to continually organize individual workplaces, however small

24

or newly established, with a majority vote. As Bryson and Gomez (2003:73) assert: “a ‘key

feature of unionization’ in decentralised majority-based union representation systems is that

‘increasing the flow of members into unions is far more difficult than maintaining the existing

stock’” (Bryson and Gomez, 2003). This statement assumes that the trade union movement, as it

stands, is incapable on its own to educate and attract workers in small firms and new industries.

This, of course, is not a given and a new strategy by unions may emerge to target these

workplaces and workers. Indeed the last section of our paper is devoted to just such possibilities.

Two caveats, however, in the experience-good/never-membership story have to be added

in order to fully account for the declines in unionisation witnessed in the private sector and the

wide disparity of trade union representation observed between the public and private sectors.

The first issue revolves around the role of employer opposition (e.g., requiring employees

to attend anti-union speeches by the employer, meeting between supervisors and small groups of

bargaining unit employees, the distribution of anti-union literature, threats against union

supporters, and promises of higher wage or benefits) in reducing support for the union and the

probability of certification. There is evidence that public sector employers do not resort to the

same degree of employer opposition as private employers do (Freeman and Kleiner, 1990).

The second issue has to do with establishment size. As seen in Figure 5, for the largest

establishments in Canada (i.e., those with 500+ employees) the incidence of unionisation is well

over 50 percent. In other words, roughly one out of every two workers in large employment

settings is currently represented by a trade union. In contrast, the level of union representation

for the smallest employers (those employing fewer than 20 employees) is 14 percent in Canada,

or roughly one in seven workers, a figure almost in line with the overall union density rate

observed in the US. This is also where the largest concentration of private sector employment is

25

housed (close to 30 percent) as compared to just over 17 percent in the public sector (see

Statistics Canada. Table 282-0075).

The average establishment size in the public sector is therefore larger than it is in the

private sector, i.e., more public sector workers are employed in large establishments than is the

case in the private sector (26 versus 12 percent respectively). There is necessarily then a

confounding relationship at work suggesting that part of the reason the public sector is more

heavily unionised has to do with a “need for voice” (arising both on the employee and employer

side) that emerges in any large establishment. And voice provision, in our labour relations

system, is necessarily met by trade union representation. These data are all the more striking in

their consistency across jurisdictions sharing decentralised workplace model of representation

(i.e., UK and the US share similar gradients in union density as firm size increases).

[Figure 5]

One implication of the data in Figure 5 is that as newer (and presumably) smaller private

sector workplaces become the norm in most Anglo-American countries, it becomes difficult and

inefficient for large national unions, from an operating standpoint, to organize these workers.

This could therefore provide a justification for a number of remedies such as allowing more

informal voice mechanisms to emerge gradually, a “broader” bargaining model and/or expanded

powers for labour relations boards to revamp and resize bargaining units. This is essentially the

argument first advanced by Willman (2001) – himself a former union organizer -- and made in

the context of why UK unions were in decline amongst small private sector establishments. A

more detailed elaboration of this idea – formalised as an extension of the Baumol “cost-disease”

argument -- has been put forth by Willman et al., (2017) and is discussed below.

26

d. The Union ‘Cost Disease’ Argument

Union organisation processes are not costless. In order to develop representation structures

around an existing membership, unions must solve two collective action problems. The first, is to

encourage employees to join unions under circumstances where the benefits of such collective

action may be available to free riders (Olson 1965; Olson and Zeckhauser, 1966). The solution to

this problem is twofold: first, selective incentives via the introduction of private goods dependent

on membership to supplement the public benefits of collective action and, second, special

conditions to coerce or form constraints to encourage membership. Where these are possible,

membership may be consistently preferable to non-membership since free riding is curtailed.

However, this sets up the second collective action problem: provision of such

mechanisms have costs, and these costs must remain lower than the benefits of union

membership to avoid the free riding problem. A central organisational problem for unions is how

to control such costs, and a central dilemma is how to balance the costs of servicing an existing

membership with those of recruiting a larger one (Willman 2004).

This is exacerbated because union organisations suffer from what Baumol and colleagues

have termed the ‘cost disease’ (see Baumol and Bowen, 1966; Baumol, 2012; Flanagan, 2012).

This affects several sectors of the economy that deliver personal services – the main examples

are healthcare, education and the performing arts – in which costs tend to rise consistently faster

than productivity and inflation, because the labour input of service delivery is difficult to replace

with capital and technology.

27

Baumol characterises these as ‘stagnant sector services’ (the stagnation referred to is in

productivity growth) and contrasts them with ‘progressive’ sectors – the examples he uses are

computers and electronics – in which technology leads to rapid increases in productivity,

reduction in real unit costs and thus the prices of goods. Costs in the ‘stagnant’ sector do not in

this approach rise primarily because of general inflation, or because of excessive wage demands.

The problem is rather relative productivity growth. Salaries in the ‘stagnant’ sector tend to rise at

broadly the same rate as those in ‘progressive’ ones, but since the percentage of total costs

represented by labour costs is falling rapidly in the latter, but maintained in the former, only the

former suffer from the ‘cost’ disease. Cost disease organisations tend to experience cost and

price rises, such that their services demand an ever higher proportion of disposable income. One

characteristic of such organisations is that unit costs increase in real terms consistently over very

long periods. A second is that prices follow.

The evidence from the longest-running and most extensively analysed dataset, from the

UK, shows that unions in the aggregate display consistent long-term real expenditure increases.

This has been documented over a period of almost a century (Webb and Webb, 1907; Roberts,

1956; Latta 1972; Willman et al. 1993; Willman et al. 2017). However, they also suffer from

pressures not to increase prices (subscriptions, membership dues) in real terms; these pressures

are primarily from competition between unions and competition with never membership. The

result is a consistent pressure on union financial reserves such that, in the UK, the trade union

movement in the aggregate is asset–poor and reserves are at all time historical lows. This affects

the ability of unions to seek membership expansion over servicing of existing members.

Put simply, what the cost-disease argument reveals is that the ‘business model’ of UK

and other Anglo-American unions is extremely fragile, pointing to the need to develop other

28

revenue streams and organisational models which avoid (or suffer less from) cost disease in

order to sustain collective voice.

e. Make or Buy Dynamics in Explaining Employer Voice Choices

Finally what role does the employer play in this discussion? We know from the analysis of

‘spurts’ in union growth that their behavior can be a contributing factor in union decline or

ascendency. The question is whether they in fact thwart unionisation by negative (anti-union) or

positive (formal provision of their own brand of non-union voice) means or some combination of

both? Before we can answer this question we need to look a little closer at the nature of voice

and its provision by either unions or employers.

Defining voice as any formal mechanism for communication between management and

workers we can distinguish voice regimes from voice types. A voice regime refers to whether a

union or the firm provides the voice mechanism for workers and a voice type refers to whether

voice is direct or representative in nature. Non-union voice can be representative (such as

statutory works councils or a joint consultative committee) or direct (such as employer led team

briefings or problem solving groups)13 whereas union provided voice is (in its current version)

always representative in nature. Voice regimes and types mix at the workplace level such as

when an existing managerial practice that provides voice to workers (an innovation circle) mixes

with a collective agreement and union presence. There is also the possibility of no voice

employers in which unions are absent and there is no formal employer provided mechanism for

engagement with employees.

13 Such mechanisms can only be said to offer voice for workers where they provide for two-way communication between workers and the employer: one-way downward communication from the employer to workers does not constitute a mechanism for worker voice.

29

If we assume that firms play a key role in the provision of workplace voice then we need

to know why some workplaces provide ‘no voice’, opt to create their own joint consultative

committees (‘make’), allow unions to form and provide voice themselves (‘buy’), or allow for

both employer-provided systems to co-exist with union voice (“dual/hedge”).

Answering this question relies on transaction cost economics and institutional theory. In

essence, we are dealing with bounded rational choices by employers who subsequently face high

switching costs once a voice regime/type is chosen. A short primer on the transaction costs

approach will be followed by an application to workplace voice.

Transaction cost economics suggests that in exchanges characterised by asset specificity,

frequency of interaction and uncertainty, choices about transaction governance structures are

required. In particular, the choice whether to make or buy, or, more accurately, own or contract.

All else equal, the more idiosyncratic the investments, the greater the frequency of interaction

(and duration of exchange) and the greater the uncertainty facing the buyer, internalising the

function rather than buying from the market will be preferred (Williamson, 1975, 1987, 2013).

The vertical integration decision by the firm is the paradigmatic example.

This choice of governance mechanism is made by parties operating under bounded

rationality, faced with the possibility of seller opportunism and operating on a risk neutral basis.

The unit of analysis is the transaction and variance in governance modes is generated by

differences in and the interaction among three actor variables: i) boundedness of rationality, ii)

trust between parties (i.e. expectation of opportunism) and iii) risk preference. Where one allows

for variance in risk preferences, as in Chiles and McMackin, 1996, one in effect shifts the unit of

analysis from ecological to cognitive, focusing directly on managerial decision making and

operating with a subjective conception of costs, i.e. as experienced by managerial decision

30

makers. Focusing on conditions at the moment of regime choice allows consideration of different

patterns among the three actor variables, but at the expense of predictive power. However, it also

allows consideration of cohort effects and switching costs.

We can read this over into the analysis of employer voice regimes in the following way.

With no idiosyncrasy, single interactions (the temporary employee paid by the piece) and no

uncertainty, the employer will not need any voice and will therefore likely be a ‘No Voice’

workplace; the classic example might be the longshore hiring hall while a contemporary one

would be Uber and its platform-economy which can recruit drivers on a daily basis and monitor

their quality through customer feedback. However, the employer not operating under these

narrow conditions and wanting voice faces a governance choice problem when seeking to

‘obtain’ a voice-producing workforce. Such an employer faces three options.

The first option is ‘making’, which involves full provision of those mechanisms which

might engender employee voice, including those perceived as legitimate by employees.

Specifically, this would involve provision of non-union voice through internal employer-made

structures (be they direct or representative). ‘Buying’ voice is the second option and would, in

extremis, involve the subcontracting out to a union of all aspects of voice provision. ‘Dual’

forms, which involve a mixture of union and non-union voice are a third possibility and might be

differentiated in terms of variance in the nature of the transaction (asset specificity, frequency

and uncertainty) or of the purchasing party (boundedness of rationality, expectation of

opportunism and risk preference).

Our central concern in this section is with explaining differences in voice regimes and

types across employers and we argue this difference can be partly explained in terms of

employer decision making under uncertainty. Such an approach can explain a central feature of

31

the evolution of voice regimes in Britain and elsewhere in Anglo-American economies– i.e., the

move towards non-union voice as part or all of workplace voice regimes.

Competition in the product market appears to encourage the shift towards non-union

voice. Traditionally, this would be interpreted as a rent issue; in competitive product markets rent

sharing possibilities disappear and the benefits of unionisation for employees diminish. But for

employers, the disappearance of a union wage premia should have made union voice more

attractive suggesting that an explanation relying on risk management is at least worth

considering – i.e., employers facing high product market risk seek to control labour supply risk

through internal voice-making decisions. This move could have been accelerated by the

hollowing out of union density itself, raising questions about unions’ ability to aggregate worker

preference and increasing the risk of “buying” over “making” voice in the period under analysis.

One theoretical conjecture is that there has been declining costs (and risks) in the

employer provision of voice (i.e., making voice) in Anglo-American economies over time, from

the paradigmatic choice of union voice in the 1960s to the non-union dominated world of the

1990s. This decline in cost of ‘making’ voice-- perhaps through the increasing availability of

Human Resource (HR) professionals and the spread of HR benchmarking available – has had the

effect of lowering the make/buy threshold for firms across all risk appetites. Supporting evidence

for this conjecture has recently been presented in a meta-analysis of union-effect studies where

the authors state that the “decline in the union effect of productivity in manufacturing,

from…positive…in Brown and Medoff (1978) to…negligible…in later studies, could reflect the

spread of “good labor practices” over time as non-union firms copied attributes of union

workplaces…[such as] seniority systems, job posting,…systems for filling vacancies rather than

relying on supervisors to promote workers, establishing formal wage scales, grievance systems

32

and mechanisms for employee voice and so on” (Doucouliagos et al., 2017:155). Indeed, as the

authors point out, many studies have documented the spread of high-commitment workplace

practices in the 1980s and 1990s, which were pioneered by unionized firms but adopted in non-

union settings. To the extent that these practices provided a productivity edge to union firms in

the 1930s through to 1970s, their spread to the non-union sector eroded the union advantage over

time (Foulke, 1981, Lewin, 1990, Colvin, 2012).

Another factor in explaining the shift away from “buy” to “make” and from

“representative” to “direct” voice is the deregulation of economies and the opening up to global

trade that increased product market risk and shifted the risk appetite of firms, thus compounding

the move away from representative union voice and towards employer-provided direct channels.

It is clear that in the private sector at least, the default option in the choice of voice regime has

shifted from union to non-union, from make to buy, from representative to direct, according to

changes in the values of risk and cost variables (assuming of course that returns were equal). The

presence of switching costs in our model, however, makes regime choice ‘sticky’ rendering

radical switching (from union to non-union and vice versa) rare. This, as we shall see, is

consistent with evidence presented in Section 4.

Our conclusions are illustrative at this point but we feel that the case for modelling

employer decision-making in order to explain voice regime and voice type evolution in Anglo-

American economies has been made. It can only be a partial one and a fuller explanation of why

representative union voice has declined and direct employer types have risen, as we have argued

above, has to incorporate employee attitudes towards unions and indeed as we saw in the “spurt

theory of growth” union policies themselves. This particularly the case as union behaviour and

organizing innovations affect the cost and risk of union voice for employers, especially in the

33

private sector where they are under increased competitive pressures and the decline in unionism

has been most felt.

4. FILLING THE VOID: WHAT HAS CHANGED IN VOICE PROVISION IN THE EARLY 21ST CENTURY?

a. The Shift from Union (Buy) to Non-Union (Make) and from Representative to Direct Voice

The section starts from the premise that there is a “market void” in the provision of collective

representation in the private sector in almost all Wagner style systems of collective bargaining,

which is to say across most of the Anglo-American world. The reason for this void, as we have

just seen, is both structural (i.e., inherent in the way decentralised bargaining systems based on

majority representation work over time in a dynamic way) and behavioural (i.e., part of the

decision-making process of individual workers and firms).

Relying on the best available dataset for charting voice over time, the Workplace

Employment Relations Survey (WERS), we display the evolution of workplace voice regimes in

Britain for the private sector since the 1980s. The US and Canada have had their versions of

WERS over the years but nothing close to the consistency of measures found in the British

survey. Where findings have a uniquely British character we shall take note.

In Table 2 Panel A, we look at voice regimes and see that the proportion of workplaces

with no voice has been roughly constant over nearly three decades (e.g., 24 percent in 1984, 22

percent in 2004 and back to 24 percent in 2011). The first three rows of Table 2 demonstrate the

growing share of non-union voice over the period 1984 to 2011. The most common form of

voice at the start of period was dual channel voice (where union and non-union voice co-exist) at

30 percent. By the end of the period, dual voice had fallen to 14 percent while non-union voice

constituted 58 percent of all private sector workplace voice regimes in Britain, up from 25

34

percent in 1984. This stands in sharp contrast to the steep decline in union-only voice (18 percent

to 3 percent between 1984 and 2011).

In Panel B of Table 2 we show the parallel shift from representative forms of voice

(irrespective of the union or non-union source) to direct forms since 1984. Direct forms of voice

which include management having regular meetings with employees, to team briefings and

problem solving groups now occupy the majority of voice provision at 43 percent in 2011 as

compared to just 9 percent with representative voice alone.

[Table 2]

While the scale of the decline in private sector union voice and representative voice has

been well documented (in the form of falling union density and representation figures), the scope

of the decline is often overlooked. By scope we mean the multiple dimensions in which union

decline has manifested itself beyond just membership decreases. In Table 3 we see that whereas

almost 60 percent of private sector workplaces had at least some union members at the start of

the period, union presence dropped dramatically by the mid-to-late-2000s. It has continued to fall

reaching 28 percent in 2011. On-site union lay representation declined continuously, falling from

38 percent in 1980 to 13 percent in 2011 (Table 3, last row), suggesting a loss of unions’ ability

to represent workers effectively even where unions continue to be recognised by the employer

(Willman and Bryson, 2009). Finally, we see a fall in union recognition across British

workplaces from 50 to 17 percent between 1984 and 2011.

[Table 3]

As anticipated by our model of employer-choice, the type of voice provided has also

changed. While many authors have accounted for the decline in the British unionisation rate

(Machin 2000) and some have even considered the rise of non-union systems as a potential cause

35

(Machin and Wood, 2005) few have considered the changing nature of non-union voice, in

particular the growth of direct forms of voice. From Table 3, we see that the decline in

representative voice was more general, extending to its non-union (employer-sponsored) forms.

For instance, the percentage of workplaces with a functioning joint consultative committee (JCC)

meeting at least once a quarter fell from 26 percent in 1984 to 16 percent by 2011. Though there

was a small increase in non-union employee representatives at the workplace, the decline in

JCC’s suggests that representative voice – in both its union and non-union forms -- suffered a

substantial decline in the private sector from 1984 to 2011.

By contrast, direct voice types have been either constant or increasing in coverage since

1984. The incidence of team briefings for example has risen by more than half (31 percent at the

start of the period rising to 51 by the end of the period). Regular meetings between employees

and senior management became more prevalent over the period 1984-1990 and have remained

common (37 percent) ever since.14 On the whole, the decline in representative voice has been

gradual in the private sector whereas the incidence of direct voice rose dramatically between the

1980s and mid 2000s and has remained high ever since.

b. Some Notable Features of Workplace Voice since the 1980s

What can we conclude from the British data?

First, that voice coverage (in all forms) is as extensive in 2011 as in 1984 or alternatively

stated, there are as many workplaces offering no voice to workers today as there were in 1984.

This may seem surprising given the extent of union decline documented in Figure 1, but this is

14 The time-series on problem-solving groups is problematic because questions are not consistent over the years. Efforts to construct a more consistent series for the period 1998-2004 suggest modest growth in their use over the period (Kersley et al., 2006: 93-94).

36

because the prevalence of voice regimes and voice types have shifted substantially over the

period in question, moving from union to non-union and from representative to direct forms of

voice. On balance voice is still being offered but non-union forms have grown to fill the void left

by declines in unionisation.

Second, in the broadest terms, our picture of the shift away from union voice begins with

employers choosing voice regimes up to the 1960s in circumstances where union-based voice

regimes were common exemplars and where there were normative and mimetic pressures to

avoid non-union only regimes. The costs of opposing unions were also higher. The risk-averse

option was dual voice. Over time, the compositional shift from manufacturing to services and

inward investments by companies with higher levels of asset specificity offered examples of

non-union voice to many employers. By the 1990s, cohort effects dominate with almost all new

entrants choosing non-union voice. Where union voice persists, as seen in Table 2, it is highly

likely to do so as part of a dual channel voice regime and very rarely on its own.

This last point is also true in the Canadian case where in the private sector, companies are

free to establish non-union labour relations approaches which facilitate (and mimic) formalized

union mechanisms for employee voice outside of a union-only structure. This is not a system that

is legally available to US employers and employees because of the peculiarities of section 8 2(d)

of the National Labor Relations Act (NLRA), which prevents most non-union forms of

workplace voice that are not certified unions (Taras, 2006). Typically, such non-union forms of

employee voice take place via employer sponsored groups and committees that may deal with

issues that are of particular interest to employers such quality, cost issues, or issues related to

improving organizational outcomes (Taras, 2006) These approaches also allow for voice on

working conditions and offer an effective platform for employee management on workplace

37

productivity and innovation related issues. An example recently in Canada is WestJet, the second

largest domestic air carrier, which saw its regional pilots reject certifying as a union in the

summer of 2015 and continue under the umbrella of the company’s own non-union labour

relations model, while its main carrier pilots voted to certify as a union in spring 2017. This

“dual” voice channel persisted until late 2017 when the regional pilots finally voted to certify.

c. Some Summary Conclusions

Explanations of union decline in Anglo-American countries have often focused on employer

opposition, macro-economic environment (high-unemployment in the 1980s to early 1990s

period and low inflation) and politics (an unfavourable legislative and social climate over the

period). These explanations are better at explaining decline in unionisation in the neo-

conservative political climate of the 1980s and early 1990s than the subsequent failure of

revitalisation in what could be described, at the very least, as a neutral organizing environment

(Kelly, 1998; Fernie and Metcalf, 2013) in which voice was still being provided to workers,

albeit in non-union ways.

We therefore argue that unionisation decline in the Anglo-American world may be part of

a broader bottom-up pattern of change in voice systems. The WERS evidence suggests that

direct voice may be replacing representative voice in the private sector, whether that

representative voice is union or employer-generated. Future work examining why direct forms of

voice have replaced representative forms could hold the key to understanding union decline. 15

15 We are not the first to tread this ground. In 1999 the late Leo Troy published Beyond Unions and Collective Bargaining and his answer to the question of why unions have declined was simple: most workers prefer individual (direct) as opposed to collective (representative) voice. The “product” the unions are selling—their services in negotiating and administering a collective labor contract—is one that, according to Troy, many workers have come to see as unattractive or irrelevant. A key reason in Troy’s model for the shift to individualised contracting is that business managers have become, on the whole, far more sophisticated in their handling of workers than they were in the era of advancing unionism. This is not a wholly unreasonable argument but if fails to consider the considerable

38

5. SOME POSSIBLE FUTURES FOR EMPLOYEE VOICE a. Traditional ‘Wagner style Unionism” Reimagined: Or why ‘training wheel’ versions of collective voice may be required With specific reference to systems that are most like the Wagner Act model (exclusive

representation systems), a potential new form of unionism that is suited to smaller workplaces

facing greater market competition and smaller profit margins is possible. We are talking about

workplaces that often are not the targets of organizing drives by unions (because of the costs

associated and the likelihood of failure) and where workers lack the knowledge and/or expertise

to undertake their own independent organizing efforts. There is also just less scope for

traditional collective bargaining over wages and benefits in such highly competitive product or

service markets because firms are in effect wage and price takers (Brown, Bryson and Forth,

2009). One might say, what else is there left to bargain over in such circumstances?

The answer is a whole lot. Everything from scheduling, procedural fairness in the

allocation of promotional or training opportunities, to simple dignity at work; all of these issues

often require a collective solution and hence representative voice to secure (Metcalf, Hansen and

Charlwood, 2001). Indeed despite having a high degree of job security and access to a non-union

employee voice system, it was scheduling concerns, not pay, that lay behind the WestJet pilots

decision to abandon their non-union association and successfully organize as a certified union

(described above).

Our reading of the historical and cross-sectional evidence, as well as the wording of

actual legislation, suggests that existing provisions under the Labour Relations Acts (LRAs) of

most jurisdictions allow for the formation of single or multiple issue employee groups that could

costs involved in organizing under the existing Wagner-Act model and does not acknowledge the consistent “latent” demand for representation found in representative surveys in the US, Canada and the UK.

39

represent all or just a fraction of the workers in a workplace. This is certainly the case in Britain

where the voluntarist character of collective representation still applies (despite a statutory

procedure available with formal certification procedures introduced during the mid-2000s under

the New Labour government) but is also true in Canada and to a lesser degree in the US.

Specifically in Ontario Sections 5 and 7 of the LRA that pertain to forming a legitimate

union and the protections members of such an association are afforded, suggests that at a

minimum any employee group that decides to form a union (which in our reading is actually not

onerous and could be aided by the provision of “boilerplate” union constitutions) would have the

protection of unjust dismissal and Labour Boards would have the authority to reinstate workers

that formed such employee associations.

In the US Section 7 of the National Labor Relations Act (NLRA) provides employees

with explicit protection in that “the right to self-organization, to form, join or assist labor

organizations, to bargain collectively through representatives of their own choosing, and to

engage in other concerted activities for the purposes of collective bargaining or other mutual aid

or protection . . .” is expressly permitted. Notably, Section 7’s wording protects both certified

union and “other concerted activity for mutual aid and protection” undertaken by non-union

workers. Stated differently, the NLRA protects employees’ exercise of their Section 7 rights

independent of whether a certified union is involved. The NLRA applies to all employee conduct

so long as employees are engaged in “concerted protected activity,” i.e., activity undertaken

together by two or more employees, or by one on behalf of others, “when they seek to improve

terms and condition of employment or otherwise improve their lot as employees.” Recent

decisions by the National Labor Relations Board (NLRB) have expanded the terms by which an

40

activity retains the protection of Section 7 to include Facebook posts and other social

communications by employees.16

To be clear, in the examples above we are talking about the formation of independent

associations of employees that could arise around a single issue – such as a hypothetical change

to benefit provision – that many employees (if not all) in a non-union setting feel was poorly

handled by management or without adequate consultation. The non-union employee group would

not necessarily be seeking “certification” in order to be granted the right to formally bargain with

their employer over an entire panoply of issues, but given that their employer is likely small or

otherwise benevolent, they would at the very least have the ability to inform management of the

collective views of their fellow employees and ask that management reconsider their decision.

Such participation systems can provide opportunities for improved communication flows

between workers and management and allow “workers to fine-tune otherwise blunt managerial

initiatives to better suit the situations of workers” (Taras, 2002).

This is an in-between model of employee representation that clearly does not afford

workers the protections afforded under the LRA certification route that if successful, compels an

employer to bargain under good faith and which in some jurisdictions afford the right to first

contact arbitration. But what it does do is reduce the costs to employees of trying to establish

some mechanism for collective voice; provides a stepping stone for workers that may one day

choose to organize and certify in the traditional way; as well as allow employers to hear from

16 Two cases in particular have expanded the role of non-union employee activity to include public shaming of employers for worker mistreatment. In Pier Sixty, LLC, 362 NLRB 59 (March 31, 2015), the NLRB affirmed an administrative law judge’s finding that the employer Pier Sixty violated Sections 8(a)(1) and (3) of the NLRA rights when it fired an employee for protected, concerted comments which were posted on his personal social media account. Similarly, in a second case the US Court of Appeals for the Second Circuit released a summary order and held that an employee’s “like” on Facebook can be protected by Section 7 of the NLRA. Three D, LLC v. NLRB, ---Fed.Appx. ----, No. 14-3284, 2015 WL 6161477 (2d Cir. Oct. 21, 2015).

41

employees who may otherwise be happy in their jobs but who have a collective grievance or

issue that they feel has been overlooked or handled poorly.

The role of statutory (employment) law in lowering the costs of ‘sampling” this kind of

representation is an important one. Looking to statutory regimes outside of the Anglo-American

sphere (in Germany and across Europe) one can see that these systems make the formation of

works councils rather easy, at least when compared to organizing a certified union in the Anglo-

American world.17 In a recent paper Forth, Bryson and George (2017) show that this legislative

support for worker representation is one of the major reasons for cross-country variance in

workplace-level worker representation across the EU. There is as a consequence a much wider

spread of representative voice coverage in regimes where the set-up costs of that voice are low.18

In short, this form of employee voice could offer a way for workers to receive

representation in a tailored and independent way, especially in the majority of private sector

establishments that are often small, made up (in many cases) of temporary and part-time workers

that have very little chance of ever being targeted by a traditional union organizing drive and/or

of becoming full blown bargaining agents under the guise of a “certified” union. Rather, this

“training wheel” model of voice is a way of potentially rebuilding the union movement around a

grass-roots, bottom-up set of employee groups that could more easily “unionise” but not

necessarily aspire to formal collective bargaining, or at least not in the first instance.

17 In one case, a large German technology company was reportedly forced to establish a works council because it met the minimum threshold of at least five employees wanting one even though 90 per cent of all other employees opposed the idea. See Allison Drutchas, “An American Works Council : Why It’s Time to Repeal NLRA Section 8 ( a )( 2 )” (2016) 93:29 Univ Detroit Mercy Law Rev 29 at 40. 18 What is interesting and has to be acknowledged is that even when the costs are pretty low the demand (either by employees or employers) is not always there. Thus the absence of works councils in smaller newer firms that are still above the threshold for works council representation.

42

b. Is the Employment Tribunal System a Model to Emulate in a World of Declining Private Sector Unionism? One of the unique features of the English-speaking world, outside of North America, most

notably in New Zealand, Ireland, Britain and Australia, is that most employment issues and

disputes are regulated by independent ‘third parties’ or ‘tribunals’ accessible to all workers

(whether unionised or not). These tribunals can make decisions with respect to workplace related

disputes and enforce certain aspects of employment law.

In New Zealand it is the Employment Relations Authority (ERA), set up under the

Employment Relations Act of 2000; in Ireland there is the Workplace Relations Commission

(WRC) formerly the Employment Appeals Tribunal which as of 2015 was folded under the

WRC; in Britain the Employment Tribunal (ET) deals with issues such as unfair dismissal,

discrimination, and unfair deductions from pay (there are similar tribunals in Northern Ireland

and Scotland); while in Australia the Fair Work Commission (FWC) deals with even a wider

remit of cases and regional agencies (such as the South Australian Employment Tribunal) that

are entrusted with resolving workplace-related disputes and individual employment issues.

Each of these are examples of an independent arbitral body with the power and authority

to regulate and enforce provisions relating to minimum wages and employment conditions for

non-union workers, and for unionised workers related to collective bargaining disputes,

industrial action, individual dispute resolution, and termination of employment.

Australia’s FWC is particularly interesting, and the focus of most of our discussion,

because it was created after a 12 year period of ‘neo-liberal hegemony’ in Australian industrial

relations (Cooper and Ellem, 2008), which weakened collective bargaining coverage and

increased the share of non-unionised workers. The FWC deals with unfair dismissal and

redundancy claims for union and non-union workers; something that in North America is not

43

covered by statutory protection. The only redress non-union workers have is through costly

Common Law court proceedings while unions deal with this in their internal grievance dispute

systems. The FWC was introduced by the Fair Work Act of 2009 and was an attempt to create a

national system for regulating industrial relations in Australia (Teicher, Holland, and Gough,

2011). It is also involved in the process of determining national industrial relations policies,

including setting minimum wages and regulating the modern award system (akin to the old

wages councils).

The FWC is similar in some respects to Britain’s Advisory, Conciliation and Arbitration

Service (ACAS) but with statutory power to have hearings and impose binding decisions.

Accordingly, it has some of the powers that the British Employment Tribunal system in that it

can resolve individual disputes but extends these to wider areas of employment and labour law.

FWC members are tenured until age 65 and the Fair Work Act of 2009 requires that those

appointed to the Commission have appropriate knowledge or experience in relevant fields. The

one drawback is that members are appointed by the Governor-General of Australia on the

recommendation of the Australian Government of the day and there is always the potential for

political bias surrounding the appointments process.

Commission Members issue binding arbitration decisions on parties and perform a quasi-

judicial role by interpreting and applying the relevant labour laws and considering evidence. The

written decisions reflect the fact that they are decisions following arbitration rather than a

judicial hearing. The Commission can also make non-binding recommendations and deal with

labour disputes. All cases are heard by a Commissioner sitting alone. Legal representation is not

a right and parties need to seek leave to be legally represented.

44

Because the FWC was designed to fill a void left by a lack of collective bargaining in

major sectors of the economy, statutory employment laws in Australia (like in most jurisdictions

with low collective bargaining rates) are extensive and complex. In Australia there are also

different states and jurisdictions with their own laws, such as in the areas of human rights and

discrimination. This is potentially confusing for employees, trade unions and employers alike as

it is not possible to raise multiple claims such as discrimination and unfair dismissal before the

FWC. A claimant has to decide at the outset what sort of claim to make and in what jurisdiction.

Once the claim has been presented it is very difficult to pursue a different basis of claim in a

different jurisdiction involving the same matter.

The Employment Tribunal system as originally established in Britain also has its merits.

Although recent changes have had the effect of lowering the numbers of workers using the

system, it is a judicial system with appropriately qualified full-time specialist adjudicators who

make judgments based on application of the law and consideration of the evidence. Parties

receive fair hearings and judgments as a consequence. It has been described as a “one stop shop”

for individual workplace disputes with jurisdiction to deal with all employment matters including

unfair dismissal, discrimination and contract claims. This makes it far more straightforward than

raising proceedings in the Common Law court system in North America. Tribunal judges in

Britain, unlike Australia, are appointed on merit, independent of the Government which means

they are not political appointees. Parties are also entitled to legal representation in an area of the

law which is increasingly complex and difficult.

Of course, this is all possible in Britain which, apart from devolution, is still effectively a

unitary state and not one with the severe geographic and federal division present in Australia or

the US and Canada for that matter. But as a model for minimal individualised “voice” and

45

redress of complaints and potential abuses, there is a lot to admire in the tribunal systems

developed outside of the North American context. Unions, who in some cases might oppose

these intrusions in what would otherwise have been rights-based grievances handled by them,

could (and have) found a role in the British and Australian tribunal systems by providing legal

and other services to employees bringing forth cases, thus providing another membership-based

model of unionism, albeit an individualised as opposed to a collective one.

c) Bespoke Unionism and ‘New’ Online Models of Worker Voice Looking to the future, as technology and artificial intelligence produce more viable substitutes

for human labour, workers especially will face competition from machines as well as from young

peers in lower income developing countries. Better training and more education can surely help,

but it is hard to imagine the average worker in the UK, Canada or the US obtaining higher pay,

benefits, and job security without also gaining a greater say in what happens at their workplace

and greater ability to defend their collective interests. Historically, trade unions have been the

major societal force for reducing inequality and raising the pay, benefits, and job security of

workers (Freeman and Medoff 1984; Blanchflower and Bryson, 2007). But while workers,

especially the young, often want to unionize, they have largely been unsuccessful with rates of

unionisation for young workers that are often one fifth that of older workers (Bryson, Gomez,

Gunderson and Meltz, 2005).

Perhaps part of the reason unions have been so unsuccessful with youth under the

traditional organizing model has less to do with ‘cost-disease’ arguments (Willman et al., 2017)

or with the prevalence of young workers located in newer workplaces (Machin, 2000), but rather

with important cohort shifts that have, hitherto, been unknown or played down. These have

46

direct implications for the potential of new models and e-platforms to form the basis of new

forms of worker voice. These include the following three new developments:

(i) The reaching of a never membership tipping point whereby the vast majority of new

labour market participants (born after 1985) have been untouched by old forms of unionisation.

This means non-unionism is the new default with direct implications for unionism as an

experience good. This makes attracting new members particularly urgent. But can unions shift

forms of reaching young workers and more generally inculcating themselves in young peoples’

lives, or do they fade even more into the background of a young person’s day-to-day existence?

(ii) The millennial population is the same generation to be fully immersed in digital

technologies. They are, in the parlance of the current age, “super-conversant” with it.

Accordingly, e-platforms become a standard, “second-nature” technology that they are willing to

access at any time for near any type of need. As documented somewhat amusingly by

millennials themselves (Harris, 2017), this is not true of older generations. The obvious question

is whether new generations will utilise this new technology to tackle workplace-based problems?

Initial signs suggest the answer is ‘yes’. The group that has obtained the most attention in

this regard is OUR Walmart, which originally adopted a Facebook-based strategy for

campaigning within Walmart and has now developed an App technology with support from IBM

Watson as a virtual expert and advisor. The App can be adapted by workers to other firms,

depending on their labor practices and policies and thus can be viewed as an experiment with

potential extensions to other firms. OUR Walmart has a miniscule staff (4 people) and

comparably small budget and by necessity operates by providing platform tools for workers to

self-organize. When it introduced its App in the fall of 2016, Walmart advised workers not to

download it – a sure sign that the firm viewed the App as a potential danger to its mode of

47

operating. OUR Walmart believes that its activities over the past two years spurred the February

2016 increase in minimum pay at Walmart, which induced its main competitors to increase their

pay commensurately. Even if OUR Walmart was only a partial factor, it is a remarkable

achievement as Walmart is the largest employer in the US, with 1.5 million “associates”, whose

labor practices influence those of other retail firms throughout the country. It is also adamantly

opposed to traditional unions. If modern technology can allow a tiny David-sized organization to

pressure such a Goliath to improve its pay and conditions, such a model of organizing could have

a huge impact on young Anglo-American workers now and into the future.

Another alt-labor group that has attracted attention is www.coworker.org, which provides

internet resources for workers to campaign for specific improvements at their workplace.

Coworker.org's design is to develop petitions that many workers sign along the lines of the

highly successful change.org. It was formed by young union activists who saw an opportunity to

help non-union workers with particular issues. Given that they and the larger change.org run

many petition campaigns on worker issues, it is perhaps an opportune time for labour relations

scholars to begin analysing the outcomes of these campaigns and identify correlates of their

success or failure.

(iii) Young workers (those under 35) are the first truly post-modern generation. They

do not take for granted the welfare state, liberal values and the like. Instead they live in a more

heavily ‘contested’ world where issues of identity and control are key. Yet it is unclear what

channels individuals can pursue to ‘fix’ their problems. Perhaps trust in leaders and politics

generally. is particularly low for the new generation and there is openness to ‘do-it-yourself’

voice. Responding to the current situation, some pro-worker activists have concluded that the

best way to help young (and other) workers is through social media and e-based platforms that

48

young person’s use to a greater extent than older persons. These forms of modern

communication and technology could potentially provide the channel for collective voice by

which workers could pressure employers for better labor conditions absent traditional unions and

collective bargaining.

Various “alt-labor” groups (Economist, 2013, Hackman, 2014; Eidelson, 2013) now

pursue the once ‘far-out notion’ of workers sharing information openly about organizing

campaigns and gaining collective power to improve their economic situation through modern

communication technology (Bryson, Gomez and Willman, 2010; Freeman and Rogers, 2002). If

firms like Uber and AirBnB can upend traditional business practices and grow suddenly through

Internet-based networked platforms (Parker et al., 2017), it seems reasonable that groups of

workers can use similar technologies to upend traditional union practices, resurrect collective

action, and improve worker well-being.

Some of these “alt-labor” groups receive traditional union support – such as Workers

Action Centres (WACs) that have led the ‘fight-for fifteen’ campaigns across many North

American cities -- but most are on their own, with few resources. They resemble the start-ups of

the dot.com period and all it takes is for at least one of them to succeed and traditional unions

will likely imitate their strategies or amalgamate/merge and transform the way they operate. The

union explosion that began in the Great Depression era was initiated by groups using the “new”

industrial union model of the CIO, which was then copied by the more ‘conservative’ AFL craft

unions. We have clearly been here before, but just as history never repeats itself in exactly the

same way, there is room for yet another voice ‘track’ to emerge which can channel the views and

interests of workers for the century ahead.

49

6. CONCLUSION This paper has reviewed the evidence on union decline in the Anglo-American world, focusing

on the connection this has had on the patterns and provision of employee voice, with a particular

emphasis on the rise of alternative forms of workplace voice and representation. Canada, Britain,

and the United States, the focal countries of our work, provide a potentially valuable set of

countries for examining the phenomenon of union decline and rise of alternative voice. All three

countries now collect comparable data on union status in their regular labor force surveys and

their collective bargaining institutions make them suitable for studying the relationship between

unions and voice provision; e.g., bargaining is highly decentralized; there are no general

mechanisms for extending collective bargaining provisions beyond the "organized" sector; and

the fraction of the private sector work force covered by collective bargaining is relatively

modest. It is therefore possible to compare the voice systems available to workers that are

covered by union contracts to those who are not covered, and potentially answer such questions

as the effect of non-union voice on overall union coverage.

Our conclusion is that trade union representation and employee voice have been on a

“twin track” for some time in all three countries. The most obvious case is the difference in

unionisation found in all major Anglo-American economies between the private and public

sectors, with the latter seemingly able to achieve a high and stable union-density coverage rate

while the latter has been in decline for over for 30 years (more in the case of the US) and

approaching historically low levels. But there was other evidence of “twin-track” unionism such

as the division found between large well-established employers (where representative union

voice still prevails) and smaller newer organizations that seem to have eschewed union and

representative forms of voice completely.

50

We found that the “spurt’ model of unionisation, first formalised by Freeman (1998), is

consistent with these trends. This model-- with its emphasis on bottom-up processes of

organizing and union and employer responses to union density level changes -- is also consistent

with other approaches such as mobilisation theory (Kelly, 1998), strategic-choice theory

(Kochan Katz and McKersie, 1983), the experience-good model of unionism (Gomez and

Gunderson, 2004) and the make-or-buy dynamics of transaction cost theory applied to employer-

provided voice choices (Willman et al., 2014).

We ended our paper with some possible futures for employee voice, having concluded

that the decline in traditional workplace-based collective bargaining is not to be taken as the

‘end’ of representative employee voice as some have suggested (Troy, 1999). The periods of

representative voice explosions (mostly union-based) have coincided with innovations in

institutional forms and new grass-roots organizing, something we contend is still possible given

the evidence over the past century in Anglo-American economies. Finally, there is also a

possible role for statutory regimes (i.e., tribunal systems or employee committees) if only to

provide some measure of voice and balance to the employment relationship while new

organizing models take hold and/or old ones are rejuvenated.

51

REFERENCES

Adams, R (2015). Bringing Canada's Wagner Act Regime into Compliance with International Human Rights Law and the Charter. Canadian Labour & Employment Law Journal, 19(1), pp.365-398.

Amosse, T., Bryson, A., Forth, J., & In Petit, H. (2016). Comparative Workplace Employment Relations: An Analysis of Practice in Britain and France. London: Palgrave MacMillan.

Baker, D. (2012). Canada Proves the Decline of Unions is not Inevitable. Retrieved from: http://www.aljazeera.com/indepth/opinion/2012/09/201291103825936625.html Barbash, J (1989). Equity as Function: Its Rise and Attrition. In Theories and Concepts in Industrial Relations, edited by Jack Barbash and Kate Barbash,.Columbia: University of South Carolina Press, pp. 114-22. Baumol, W.J. (2012) The Cost Disease: Why Computers get Cheaper and Health Care Doesn’t. New Haven: Yale University Press. Baumol, W. J., & Bowen, W. G. (1966). Performing Arts, The Economic Dilemma: A Study of Problems Common to Theater, Opera, Music, and Dance. New York: Twentieth Century Fund. Beaumont, P.B. (1986). Management Opposition to Union Organisation: Researching the Indicators. Employee Relations, 8(5), pp. 31-38. Beaumont, P.B., & Harris, R. I. D. (1994). Opposition to unions in the non-union sector in Britain. International Journal of Human Resource Management, 5(2), pp. 457-471. Blanchflower, D. and Bryson, A. (2007). What Effect Do Unions Have on Wages Now and Would "What Do Unions Do?" Be Surprised. In James T. Bennett and Bruce E. Kaufman (eds.), What Do Unions Do? A Twenty-Year Perspective, pp. 79-113, New Brunswick: Transaction Publishers. Booth, J. E., Budd, J. W., & Munday, K. M. (2010). Never Say Never? Uncovering the Never‐Unionized in the United States. British Journal of Industrial Relations, 48(1), pp. 26-52. Braley-Rattai, A. (2013). Harnessing the Possibilities of Minority Unionism in Canada. Labor Studies Journal, 38(4), pp. 321 – 340. Brown, C., & Medoff, J. (1978). Trade Unions in the Production Process. Journal of Political Economy, 86(3), 355-378. Brown, W., Bryson, A. and Forth, J. (2009) Competition and the Retreat from Collective Bargaining. Chapter 2 in W. Brown, A. Bryson, J. Forth and K. Whitfield (eds.) The Evolution of the Modern Workplace, pp. 22-47, Cambridge University Press

52

Bryson, A., & J. Forth (2010). Trade Union Membership and Influence 1999-2009. CEP Discussion Paper No 1003. Centre for Economic Performance, London School of Economics and Political Science. Bryson, A., & R. Gomez. (2003). Buying Into Union Membership. In Representing Workers: Trade Union Recognition and Membership in Britain, edited by Howard Gospel and Stephen Wood, pp. 72–91. London: Routledge. Bryson, A., & Gomez, R. (2005). Why Have Workers Stopped Joining Unions? The Rise in Never‐Membership in Britain. British Journal of Industrial Relations, 43(1), pp. 67-92. Bryson, A., Gomez, R., Gunderson, M., & Meltz, N. (2005). Youth-adult differences in the demand for unionization: Are American, British, and Canadian workers all that different?.Journal of Labor Research, 26(1), pp. 155-167. Bryson, A., Gomez, R. and Willman, P. (2010). Online Social Networking and Trade Union Membership: What the Facebook phenomenon truly means for labor organizers. Labor History. 51(1), pp.41-53 Bryson, A., & Freeman, R. B. (2013). Employee Perceptions of Working Conditions and the Desire for Worker Representation in Britain and the US. Journal of Labor Research, 34(1), pp. 1-29. Budd J.W. (2004). Employment with a Human Face: Balancing Efficiency, Equity, and Voice. Ithaca: Cornell University Press. Budd, J. W. (2010). When do US Workers First experience unionization? Implications for revitalizing the labor movement. Industrial Relations: A Journal of Economy and Society, 49(2), pp. 209-225 Commons, J.R. (1935). Communism and Collective Democracy. American Economic Review, 25(2), pp. 212-23. Cooper, R. & Ellem, B. (2008). The Neoliberal State, Trade Unions and Collective Bargaining in Australia. British Journal of Industrial Relations, 46(3), pp. 532–554 Costa, M., Hearn, M., & Lloyd Ross Forum (eds). (1997). Reforming Australia's Unions: Insights from Southland Magazine. Sydney: Federation Press. Derber, M. (1967). Research on Labor Problems in the United States. New York: Random House. Derber, M. (1970). The American Idea of Industrial Democracy, 1865–1965. Urbana: University of Illinois Press.

53

Diamond W.J.,& Freeman R.B. (2001). Liking the Workplace You Have: The Incumbency Effect in Preferences toward Unions. Working Paper 115, Centre for Economic Performance, London School of Economics. Doorey, D.J. (2013). Graduated Freedom of Association: Worker Voice Beyond the Wagner Model. Queen's Law Journal, 38(2), pp. 511-521 Doucouliagos, H., Freeman, R.B, & Laroche, P. (2017). The Economics of Trade Unions: A Study of a Research Field and its Findings. London: Routledge, Taylor & Francis Group. Dunlop, J. T. (1944). Wage determination under trade unions. New York: Macmillan Co. Dunlop J.T. (1958). Industrial Relations Systems. New York: Holt & Co. Economist, The. (2013) New labour, Alt-labour. Retrieved from: http://www.economist.com/news/unitedstates/21586287-believe-it-or-not-union-movement-starting-embrace-innovation-new-labour Eidelson, J. (2013). Alt-Labor. The American Prospect, (January 29, 2013). Retrieved from: http://prospect.org/article/alt-labor Erickson, C. L., Fisk, C. L., Milkman, R., Mitchell, D. J., & Wong, K. (2002). Justice for janitors in Los Angeles: Lessons from three rounds of negotiations. British Journal of Industrial Relations, 40(3), pp. 543-567. Feldman, H. (1928). Survey of Research in the Field of Industrial Relations. New York: Social Science Research Council. Ferguson, J.-P. (2008). The Eyes of the Needles: A Sequential Model of Union Organizing Drives, 1999-2004. Industrial and Labor Relations Review, 62(1), pp. 3-21. Fernie, S., & Metcalf, D. (2013). Trade Unions: Resurgence or Demise?. Hoboken: Taylor and Francis. Flanagan, R.J. (2012). The Perilous Life of Symphony Orchestras. New Haven, Yale University Press. Forth, J., Bryson, A. and George, A. (2017) Explaining Cross-National Variation in Workplace Employee Representation, European Journal of Industrial Relations, 23(4), pp. 415-433 Freeman, R.B. & J. Medoff. (1984). What Do Unions Do? New York: Basic Books. Freeman, R. B., & Ichniowski, C. (1988). When Public Sector Workers Unionize. Chicago: University of Chicago Press.

54

Freeman, R. B., & Kleiner, M. M. (1990). Employer Behavior in the Face of Union Organizing Drives. Industrial and Labor Relations Review, 43(4), pp. 351-365. Freeman, Richard B, and Joel Rogers. (1999). What Workers Want. Ithaca, N.Y.: ILR Press. Freeman, R and Joel Rogers (2002). Open Source Unionism: Beyond Exclusive Collective Bargaining. WorkingUSA: The Journal of Labor and Society, 7(2), pp 3-4 Freeman, R.B. (1998). "Spurts in Union Growth: Defining Moments and Social Processes". In The Defining Moment: The Great Depression and the American Economy in the Twentieth Century. Edited by Michael D.Bordo,Claudia Goldin, and Eugene N.White, Chicago: The University of Chicago Press, pp.265-295. Gomez, R., & Gunderson, M. (2004). “The Experience Good Model of Trade Union Membership.” In The Changing Role of Unions: New Forms of Representation, edited by Phanindra V.Wunnava, pp. 92–112. Armonk, NY: M.E. Sharpe. Gomez, R., Gunderson, M., & Meltz, N. (2002). Comparing Youth and Adult Desire for Unionization in Canada. British Journal of Industrial Relations, 40(3), pp. 519-542 Hackman, R. (2014). Alt-labor: a new union movement or the same old song? The Guardian (May 1, 2014). Retrieved from: https://www.theguardian.com/money/2014/may/01/low-wage-new-unionwalmart-fast-food Harris, M (2017). Kids These Days: Human Capital and the Making of Millennials. New York: Little, Brown and Company. Hirsch, B., Macpherson, D., & Vroman, W. (2001). Estimates of Union Density by State. Monthly Labor Review, 124(7), pp. 51-55. Ichniowski, C., & Freeman, R. B. (1988). When public sector workers unionize. Chicago, Ill: University of Chicago Press. Jacoby, S.M. (1985). Employing Bureaucracy: Managers, Unions, and the Transformation of Work in the Twentieth Century. New York: Columbia University Press. Jacoby, S. M. (2004). Employing bureaucracy: Managers, unions, and the transformation of work in the 20th century. Psychology Press. Johnson S (2002). Card Check or Mandatory Representation Vote? How the type of union recognition procedure affects union certification success? The Economic Journal, 112(2), pp. 344–361. Johnson S (2004). The Impact of Mandatory Votes on the Canada–US Union Density gap: A note. Industrial Relations, 43 (2), pp. 356–363.

55

Kaufman, B. E. (2010). The Theoretical Foundation of Industrial Relations and its Implications for Labor Economics and Human Resource Management. Industrial and Labor Relations Review, 64(1), pp. 74-108. Kelly, J.(1998). Rethinking Industrial Relations: Mobilisation, Collectivism and Long Waves. London and New York: Routledge. Kleiner, M., & Krueger, A. (2013). Analyzing the Extent and Influence of Occupational Licensing on the Labor Market. Journal of Labor Economics, 31(2), S173-S202. Latta, G. (1972). Trade Union Finances. British Journal of Industrial Relations, (10(2), pp.392-411 Logan, J. (2006). The Union Avoidance Industry in the United States. British Journal of Industrial Relations, 44(4), pp. 651-675. Machin, S. (2000). Union Decline in Britain. British Journal of Industrial Relations, 38(4), pp. 631-645. Machin, S., & Wood, S. (2005). Human resource management as a substitute for trade unions in British workplaces. Industrial and Labor Relations Review, 58(2),pp. 201-218. Meltz, N. M. (1989). “Industrial Relations: Balancing Efficiency and Equity.” In Jack Barbash and Kate Barbash, eds., Theories and Concepts in Comparative Industrial Relations. Columbia: University of South Carolina Press, pp. 109-13. Metcalf, D., Hansen, K., & Charlwood, A. (2001). Unions and the Sword of Justice: Unions and Pay Systems, Pay Inequality, Pay Discrimination and Low Pay. National Institute Economic Review, 176(1),pp. 61-75. Olson, M. (1965). The Logic of Collective Action; Public Goods and the Theory of Groups. Cambridge: Harvard University Press. Olson, M., & Zeckhauser, R. (1966). An Economic Theory of Alliances. The Review of Economics and Statistics, 48(3), pp. 266-279. Riddell, C. (2001). Union Suppression and Certification Success. Canadian Journal of Economics/Revue canadienne d'économique, 34(2), pp. 396-410. Roberts, B. C. (1956). Trade Union Government and Administration. London: Bell and Sons. Rose, J. B., & Chaison, G. N. (1996). Linking Union Density and Union Effectiveness: The North American experience. Industrial Relations: A Journal of Economy and Society, 35(1), pp. 78-105

56

Parker, G., Van, A. M., & Choudary, S. P. (2017). Platform Revolution: How networked markets are transforming the economy - and how to make them work for you. New York : W.W. Norton Slinn, S. (2008). No right (to organize) without a remedy: Evidence and consequences of the failure to provide compensatory remedies for unfair labour practices in British Columbia. McGill Law Journal, 53(4), pp.687-736 Taras, D. G. (2002). Alternative Forms of Employee Representation and Labour Policy. Canadian Public Policy. 28(2): pp.105–116. Taras, D. G. (2006). Non-union Representation and Employer intent: How Canadian Courts and Labour Boards Determine the Legal Status of Non-union Plans. Socio-Economic Review. 4(2): pp.321-336. Teicher, J., Holland, P., & Gough, R. (Eds.). (2013). Australian Workplace Relations. Cambridge University Press. Thomason, T. and S. Pozzebon, (1998). Managerial Opposition to Union Certification in Quebec and Ontario. Relations industrielles / Industrial Relations, 53(4), pp. 750-771. Thompson P. & K. Newsome (2004). Labor Process Theory, Work, and the Employment Relation. In Kaufman Bruce E., ed., Theoretical Perspectives on Work and the Employment Relation, pp. 133–62. Champaign, Industrial Relations Research Association. Troy, L. (2000). Beyond Unions and Collective Bargaining. Working USA, 3(5): 102–134 Webb, Sidney and Beatrice Webb. 1897. Industrial Democracy. London: Longman Greens. Webb, S. and Webb, B. 1907. I. London: Longman Green and Co. Weiler, P. C. (1989). Governing the Workplace. Cambridge: Harvard University Press. Williamson, O. E. (1975). Markets and hierarchies: Analysis and Antitrust Implications : A study in the economics of internal organization. New York: Free Press. Williamson, O. E. (1987). The Economic Institutions of Capitalism: Firms, markets, relational contracting. New York: The Free Press. Williamson, O. E. (2013). The Transaction Cost Economics Project: The Theory and Practice of the Governance of Contractual Relations. Cheltenham: Edward Elgar Press. Willman, P., Morris, T. and Aston, B. (1993). Union Business; Trade Union Organisation and Financial Reform in the Thatcher Years. Cambridge: Cambridge University Press Willman, P., Bryson, A., Gomez, R. & Kretschmer, T. (2014) "Employee voice and the transaction cost economics project", Chapter 4 In A. Wilkinson, J. Donaghey, T. Dundon, R.

57

Freeman (eds.) Handbook of Research on Employee Voice, pp. 52-65, Cheltenham: Edward Elgar. Willman, P., Bryson, A. and Forth, J. (2017). UK Unions, Collective Action and the Cost Disease. Department of Management, LSE. Willman, P. (2004). Structuring Unions: The Administrative Rationality of Collective Action. In Union Organisation and Activity, London: Routledge, edited by J. E. Kelly & Paul Willman. London: Routledge. pp. 73-85. Willman, P. (2001). The Viability of Trade Union Organization: A Bargaining Unit Analysis. British Journal of Industrial Relations, 39(1), pp.97-117.

58

Figure 1: Union Density in Six English Speaking Countries, 1970-2014

Source: OECD, accessed on OECD.stat June 7 2017.

Australia

Canada

Ireland

New Zealand

United Kingdom

United States

OECD average

0

10

20

30

40

50

60

70

Uni

on D

ensit

y (%

)

59

Figure 2: Union Density (%) in Public versus Private Sector, 1970-2015 a) United States

b) Canada

c) United Kingdom

Source: Bureau of Labor Statistics (BLS); Statistics Canada (StatsCanada); and Office of National Statistics (ONS)

0

5

10

15

20

25

30

35

40

1970 1980 1990 2000 2010 2015

Public

Private

0

10

20

30

40

50

60

70

80

1970 1980 1990 2000 2010 2015

PublicPrivate

0102030405060708090

1970 1980 1990 2000 2010 2015

PublicPrivate

60

Figure 3: Current, Ex and Never Union Membership Rates in Britain, 1986-2008

Source: Adapted from Forth and Bryson (2010), Figure 4.

0

10

20

30

40

50

60

1985-86 2006-08

Perc

ent

Current member

Ex-member

Never-member

61

Figure 4: Union Density and Timing of U.S. State-level Right-to-Work Laws, 1964-2017

Source: Barry T. Hirsch, David A. Macpherson, and Wayne G. Vroman, “Estimates of Union Density by State,” Monthly Labor Review, Vol. 124, No. 7, July 2001, pp. 51-55. Accompanying data for 2001 to 2016 online at www.unionstats.com . Notes: Figures represent the percentage of each state's non-agricultural wage and salary employees who are union members. The vertical lines indicate the year when the state “right-to-work law” was passed outlawing union shop (which allows for hiring non-union employees, provided that the employees then join the union and pay dues within a certain period) and agency shop (in which employees must pay the equivalent of the cost of union representation, but need not formally join the union) provisions in their jurisdictions.

KY

IN

MI

MO

WV

WI

All States

OK

LA

ID

TX

0

5

10

15

20

25

30

35

40

45

50

1964 1974 1984 1994 2004 2012 2016

Perc

ent U

nion

Mem

bers

(%)

Michigan (2012)Indiana (2012)

Wisconsin (2015)

Oklahoma (2001)

West Virgimia (2016)

Kentucky (2017)Missouri (2017)

Idaho (1985)

Louisiana (1976)

Texas (1993)

62

Figure 5: Union Density by Size of Establishment in Canada, 1997-2016

Source: Statistics Canada. Table 282-0224 - Labour Force Survey estimates (LFS), employees by union status, and establishment size, Canada, annual (persons unless otherwise noted).

Notes: Union density is calculated by members of a union and / or covered by a collective agreement as a proportion of all employees. Beginning January 1997, the number of employees at the location of employment (for example, building or compound) is collected from employees. Responses are recorded according to the following size groups: less than 20, 20 to 99, 100 to 500, more than 500. The concept of location of employment approximates the concept of establishment used by many Statistics Canada business surveys.

Less than 20 employees

20 to 99 employees

100 to 500 employees

More than 500 employees

0

10

20

30

40

50

60

70

Perc

ent U

nion

Mem

bers

(%)

63

Table 1: Union Density and Timing of State Right-to-Work (RTW) Laws, 1964-2017 Right-to-Work Law Union Density State Year passed Union Density High (year) Low (year) Louisiana 1976 17.3 18.1 (1964) 4.2 (2016) Idaho 1985 12.2 25.0 (1968) 4.7 (2013) Texas 1993 7.6 13.8 (1970) 4.0 (2016) Oklahoma 2001 8.5 16.8 (1970) 5.4 (2016) Michigan 2012 16.8 44.8 (1964) 14.6 (2016) Indiana 2012 9.2 40.9 (1964) 9.2 (2012) Wisconsin 2015 8.4 34.0 (1964) 8.2 (2016) West Virginia 2016 11.9 36.5 (1964) 10.6 (2014) Kentucky 2017 11.2 25.4 (1970) 8.7 (2008) Missouri 2017 9.7 27.1 (1964) 8.4 (2014) Overall Average 2004 11.3 28.2 (1966) 7.8 (2014)

Source: Barry T. Hirsch, David A. Macpherson, and Wayne G. Vroman, “Estimates of Union Density by State,” Monthly Labor Review, Vol. 124, No. 7, July 2001, pp. 51-55. Accompanying data for 2001 to 2016 online at www.unionstats.com .

64

Table 2: Prevalence of Voice in British Private Sector Workplaces, 1984-2011 1984

(%) 2004 (%)

2011 (%)

Percentage Point Change

Panel A: Voice Regimes Union voice only

18

5

3

-15

Dual voice †

30

17

14

-16

Non-union voice only

25

55

58

+33

No Voice 24 22 24 No change

Panel B: Voice Types Representative voice only 23 11 9 -14 Representative + Direct voice 35 26 24 -11 Direct voice only 17 40 43 +26 No voice 24 22 24 No change

Notes: This voice typology is constructed using the voice items which are present in the data throughout the period 1984-2011 for private sector workplaces with more than 25 employees. All values are column percentages. Columns may not add up to total voice percentages due to rounding and missing information on some workplace. † Dual voice refers to presence of union and non-union voice systems in the same workplace. Source: Authors calculations. WERS survey various waves.

65

Table 3: Prevalence of Voice Types in British Private Sector Workplaces, 1984-2011 1984

(%) 2004 (%)

2011 (%)

Percentage Point Change

Panel A: Non-Union Voice Representative Voice Join Consultative Committee (JCC)

26

20

21

-5 JCC that meets regularly

24

15

16

-8

Non-union employee representative

10†

16

15

+5

Direct Voice Regular meetings with employees 34 37 37 +3 Team briefings 31 48 51 +20 Problem-solving groups† 30 28 25 -5 Panel B: Union Voice Any union members 60 35 28 -32 Any recognized union 50 22 17 -33 Any on-site lay union representative 38 14 13 -25

Notes: This voice typology is constructed using the voice items which are present in the data throughout the period 1984-2011 for private sector workplaces with more than 25 employees. All values are column percentages. Columns may not add up to total voice percentages due to rounding and missing information on some workplace. † This question was only asked starting in 1990. Source: Authors calculations. WERS survey various waves.